AMP/ASCO/CAP Somatic Variant Tier Classification

Why use it

Reporting a tumour variant as "pathogenic" tells an oncologist nothing about whether to act on it, and reporting it with an ACMG germline tier invites exactly the wrong inference. The AMP/ASCO/CAP tier is the vocabulary molecular tumour boards, CAP-accredited laboratories and treatment guidelines actually use, and it forces the tumour-type context to be stated rather than assumed.

About

The joint AMP/ASCO/CAP standard classifies somatic sequence variants in cancer into four tiers by their clinical significance for the tumour type being tested. Tier I holds variants of strong clinical significance: those with an FDA-approved therapy or a place in professional guidelines for this tumour type (level A), or supported by well-powered studies with expert consensus (level B). Tier II holds variants of potential clinical significance: an approved therapy or guideline in a different tumour type, or investigational agents with multiple small studies and some consensus (level C), or preclinical work and isolated case reports (level D). Tier III is variants of unknown clinical significance — not in population databases at significant frequency, not reported in cancer databases, and without convincing published evidence. Tier IV is benign or likely benign variation. Critically, the tier describes actionability in a context, not a property of the variant: the same BRAF or KRAS change moves between tiers as the tumour type and the evidence base change.

When to use

• Reporting somatic variants from tumour sequencing panels, exomes or genomes
• Preparing a case for a molecular tumour board
• Deciding whether a variant supports an on-label therapy, an off-label option or a clinical trial referral
• Re-tiering an archived tumour report after a new approval or guideline changes the evidence level
• Separating somatic actionability from germline pathogenicity when a variant is found in both settings
• Teaching why a somatic report and a germline report answer different questions

Formula

Pearls & pitfalls

• The tier is tumour-type specific. BRAF p.Val600Glu is Tier I in melanoma and Tier II in colorectal cancer — the variant is identical, the evidence is not.
• Level C is the level that most often gets over-read. An approval in a different tumour type is potential significance, not strong significance, and reporting it as Tier I invites an off-label request the evidence does not support.
• Tier III means "we do not know", not "it is benign". Tier IV is the benign tier. Collapsing the two loses the variants most worth re-reviewing later.
• A high variant allele fraction near 50% (or near 100% with loss of heterozygosity) in a cancer-predisposition gene should raise the question of germline origin — but a tumour-only test cannot answer it. Confirm on a normal sample.
• Population frequency arguments work differently here. A variant common in gnomAD is Tier IV as a somatic driver, but for tumour-only sequencing without a matched normal, common variants are also the ones most likely to be unfiltered germline background.
• Tiering is not the same as variant calling confidence. A Tier I call from a 3% allele fraction with 40× coverage still needs orthogonal confirmation before it changes treatment.
• The 2017 standard predates several tumour-agnostic approvals. A tumour-agnostic indication (MSI-H, TMB-high, NTRK fusion) is level A evidence in any tumour type that meets the biomarker definition — do not downgrade it to level C merely because the trial cohort was another histology.
• Record the knowledge bases and their versions. OncoKB, CIViC and the NCCN guidelines change on a timescale of months, and a tier without a date is not reproducible.

Next steps

Tier I: report prominently — the variant is actionable in this tumour type, either therapeutically, diagnostically or prognostically. Tier II: report with the evidence context stated explicitly, including whether the supporting approval is in a different tumour type; consider clinical trial matching. Tier III: report in a separate section as being of unknown significance, with no treatment implication, and re-review as the evidence base grows. Tier IV: generally not reported. In every case state the tumour type the tiering applies to, and if a germline origin is suspected, arrange confirmatory germline testing rather than inferring it from the tumour result.

Evidence

Li MM, Datto M, Duncavage EJ, et al. Standards and guidelines for the interpretation and reporting of sequence variants in cancer: a joint consensus recommendation of the Association for Molecular Pathology, American Society of Clinical Oncology, and College of American Pathologists. J Mol Diagn. 2017;19(1):4-23. The four-tier structure and the A-D evidence levels follow that standard directly.

References

1. Li MM et al. Standards and guidelines for the interpretation and reporting of sequence variants in cancer: a joint consensus recommendation of AMP, ASCO and CAP. J Mol Diagn. 2017;19(1):4-23. — link
2. Horak P et al. Standards for the classification of pathogenicity of somatic variants in cancer (oncogenicity): joint recommendations of ClinGen, CGC and VICC. Genet Med. 2022;24(5):986-998. — link
3. Chakravarty D et al. OncoKB: a precision oncology knowledge base. JCO Precis Oncol. 2017;1:1-16. — link